BERTraffic: BERT-based Joint Speaker Role and Speaker Change Detection for Air Traffic Control Communications
This work addresses a domain-specific challenge in air traffic control for improving automatic speech recognition and entity extraction, but it is incremental as it builds on existing BERT and speaker diarization methods.
The paper tackled the problem of speaker change and role detection in air traffic control communications by proposing a BERT-based system that combines speech activity detection and BERT models, achieving up to 10% and 20% token-based Jaccard error rates and relative improvements of 32% and 7.7% in error rates compared to a baseline system.
Automatic speech recognition (ASR) allows transcribing the communications between air traffic controllers (ATCOs) and aircraft pilots. The transcriptions are used later to extract ATC named entities, e.g., aircraft callsigns. One common challenge is speech activity detection (SAD) and speaker diarization (SD). In the failure condition, two or more segments remain in the same recording, jeopardizing the overall performance. We propose a system that combines SAD and a BERT model to perform speaker change detection and speaker role detection (SRD) by chunking ASR transcripts, i.e., SD with a defined number of speakers together with SRD. The proposed model is evaluated on real-life public ATC databases. Our BERT SD model baseline reaches up to 10% and 20% token-based Jaccard error rate (JER) in public and private ATC databases. We also achieved relative improvements of 32% and 7.7% in JERs and SD error rate (DER), respectively, compared to VBx, a well-known SD system.